142, METEOROLOGICAL OBSERVATIONS. 
cumftances too attending the barometer I thought deferved- 
to be particularly noted, which could not have been done 
had the firft idea been adopted. For the obfervations 
upon the barometer not only fhew us the different 
ftates of the atmofphere, but, perhaps, may throw far- 
ther light upon the true caufe of the Aurora Porealis. 
The fact is, that a fall of the barometer always fucceeds 
that phenomenon. The frequency of its appearance 
lately, gave me an opportunity of obferving this effect 
at different times. It has for fome time been fuppofed 
(after Dr. Franklin had firft given rife to the opinion) 
to be an electrical appearance ; and I think, the levity of 
the atmofphere, as proved by the barometer, adds great 
weight to that fuppofition: fince it is well known to every 
electrician, that a rarefaction of the air, in our experi- 
ments, will always produce fimilar appearances. One 
circumftance indeed was obfervable, that a change of wea- 
ther, to wet, generally fucceeded; but as this effect was not 
fo conftant, it was not much attended to. But the baro-— 
meter by fhewing that the atmofphere is aétually lighter, 
and of confequence more rarefied at the time of fuch ap- 
pearances than at others, evinces at leaft, that itis ina 
ftate the moft likely to exhibit them; it is to be obferved 
alfo, that the greateft fall of the barometer is not prior to, 
but always fucceeds this appearance; fhewing that the ra- 
refaétion firft begins in the upper parts of the atmofphere. 
It is remarkable that the range of the barometer was 
not more than one inch and a tenth throughout the whole 
year, nor do I remember ever to have feen a greater dif- 
ference at any time not included in the journal ; whilft we 
fee in other countries, the atmofphere undergoing changes 
fo great as to effect a difference of three or four inches. 
Whence is it then that we are expofed to more violent 
ftorms of wind and rain? Perhaps indeed the changes here, 
though not fo great, may be more fudden, of which fome 
remarkable inftances may be feen in the journal. 
Our 
